action times out on latest golangci-lint version

[pauleibye]

Golangci-lint just had a new release, 1.47.0, which broke our lint action. The action would spin and timeout when running the golangci-lint command with or without args.

Example log where the action would hang:
Running [/home/runner/golangci-lint-1.47.0-linux-amd64/golangci-lint run --out-format=github-actions --new-from-patch=/tmp/tmp-1669-7BzhRHngF7zi/pull.patch --new=false --new-from-rev=] in [] ...

Downgrading the golangci-lint version to 1.46.2 fixed the action for us.

[iwata]

I'm having the same problem too.

[sbueringer]

I have the same issue.

Not sure if it's specific to my case but when I run the action locally with https://github.com/nektos/act I can reproduce the issue. When I exec into the action and run golangci-lint I run into the following panic:

docker exec -it 13d3 /bin/bash
root@sbuerin:/home/sbuerin/code/src/sigs.k8s.io/cluster-api# /root/golangci-lint-1.47.0-linux-amd64/golangci-lint
panic: load embedded ruleguard rules: rules/rules.go:13: can't load fmt

goroutine 1 [running]:
github.com/go-critic/go-critic/checkers.init.22()
        github.com/go-critic/[email protected]/checkers/embedded_rules.go:47 +0x4b4

(The binary seems to just panic in general, independent of in which folder it is called, Go 1.17 or Go 1.18, gocritic enabled or disabled)

P.S. I reproduced this with kubernetes-sigs/cluster-api#6943 + act -j golangci

[TenSt]

I have the same issue. In my case, the action would just run without stopping and I need to stop them manually. All works fine on v1.46.2, so we're downgrading for the time being too.

[ldez]

If you have the same kind of problem, I need:

• your Go version
• your golangci-lint configuration
• the action configuration
• a public repo (if possible)

[sbueringer]

• Go 1.17.11
• https://github.com/kubernetes-sigs/cluster-api/blob/5d60581784610d7c7c8350258740f7fab5aa570b/.golangci.yml
• https://github.com/kubernetes-sigs/cluster-api/blob/5d60581784610d7c7c8350258740f7fab5aa570b/.github/workflows/golangci-lint.yml
• https://github.com/sbueringer/cluster-api/tree/pr-bump-golangci-lint

I could reproduce by:

• checking out this PR: 🌱 Bump golangci-lint to v1.47.1 kubernetes-sigs/cluster-api#6943
• Running act -j golangci (https://github.com/nektos/act)

[TenSt]

@ldez, here it is:

• Go 1.18.x
• we use the default one
• job config:

go-lint:
  name: Lint Go Applications
  if: needs.vars.outputs.go_changes == 'true'
  runs-on: ubuntu-latest
  needs:
    - vars

  steps:
    - uses: actions/checkout@v3

    - uses: actions/setup-go@v3
      with:
        go-version: "1.18.x"
        cache: true
        cache-dependency-path: go/go.sum

    - name: golangci-lint
      uses: golangci/[email protected]
      with:
        working-directory: go
        version: latest
        args: --timeout 5m

• the repo is private, so I can't share

[flemzord]

I have a same problem in my github action

[sbueringer]

I bumped to Golangci lint v1.47.1. It looks better now.

Previously it timed out after 6 hours, now I'm hitting my own timeout of 10 minutes (action times out after 15m).

We are bumping from v1.44=>v1.47.1 and will try to figure out if it's expected that the linter now takes ~ 15 min instead of 6-7 min before.

[ldez]

I think the problem is solved, for the performance issue you can follow golangci/golangci-lint#2997